Google Messages adds AI-detection to every chat photo

That suspiciously perfect photo in your group chat might finally come with a paper trail. Google Messages has rolled out support for Content Credentials, letting users check the editing history behind photos, videos, and audio files shared in conversations, including whether AI tools touched them along the way.

The feature is built on C2PA standards, which is a system used to track how media files have been generated and altered. It serves as the paper trail for the media file – like in the example above, the sunset photo would show that it originated from a camera photo but later was altered using AI software.

When Messages has detected that the media file has been edited but cannot determine if it has been done using AI software, the file is marked as being "edited using AI tools".

The feature has a hard limit worth remembering: it can't confirm whether an image is real or fake, only what history is attached to the file. A photo with no content credentials at all isn't necessarily untouched.

It may just be a case where the file went through an application that does not support the new format; hence, no trace of its existence does not mean anything about its authenticity.

The tool currently only works on Android devices and does not apply to the Google Messages Web version. In order to use it, press and hold the content in question, click info, then scroll down to How This Was Made.

If Content Credentials is available for that particular content, Messages will indicate whether it was originally created by a camera or a microphone and if there was any editing using either AI or non-AI tools.

Content credentials won't settle every dispute over a questionable photo just yet. Wide adoption of the standard across cameras, apps, and platforms is still catching up, meaning plenty of media people send still won't carry any record at all.